From 7eac52648a4c24ad23a05f62db97867c92a5747b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 7 Feb 2020 19:11:12 -0500 Subject: SUNRPC: Add a flag to avoid reference counts on credentials Add a flag to signal to the RPC layer that the credential is already pinned for the duration of the RPC call. Signed-off-by: Trond Myklebust --- include/linux/sunrpc/sched.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/sunrpc/sched.h b/include/linux/sunrpc/sched.h index a6ef35184ef1..df696efdd675 100644 --- a/include/linux/sunrpc/sched.h +++ b/include/linux/sunrpc/sched.h @@ -132,6 +132,7 @@ struct rpc_task_setup { #define RPC_TASK_TIMEOUT 0x1000 /* fail with ETIMEDOUT on timeout */ #define RPC_TASK_NOCONNECT 0x2000 /* return ENOTCONN if not connected */ #define RPC_TASK_NO_RETRANS_TIMEOUT 0x4000 /* wait forever for a reply */ +#define RPC_TASK_CRED_NOREF 0x8000 /* No refcount on the credential */ #define RPC_IS_ASYNC(t) ((t)->tk_flags & RPC_TASK_ASYNC) #define RPC_IS_SWAPPER(t) ((t)->tk_flags & RPC_TASK_SWAPPER) -- cgit v1.2.3 From 8d6bda7f23a9b3ef1d7e386f01924c37f18fe771 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Wed, 11 Mar 2020 11:21:12 -0400 Subject: SUNRPC: Remove xdr_buf_read_mic() Clean up: this function is no longer used. Signed-off-by: Chuck Lever Reviewed-by: Benjamin Coddington Signed-off-by: Trond Myklebust --- include/linux/sunrpc/xdr.h | 1 - 1 file changed, 1 deletion(-) (limited to 'include') diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h index b41f34977995..8a6dd5bd6748 100644 --- a/include/linux/sunrpc/xdr.h +++ b/include/linux/sunrpc/xdr.h @@ -184,7 +184,6 @@ xdr_adjust_iovec(struct kvec *iov, __be32 *p) extern void xdr_shift_buf(struct xdr_buf *, size_t); extern void xdr_buf_from_iov(struct kvec *, struct xdr_buf *); extern int xdr_buf_subsegment(struct xdr_buf *, struct xdr_buf *, unsigned int, unsigned int); -extern int xdr_buf_read_mic(struct xdr_buf *, struct xdr_netobj *, unsigned int); extern int read_bytes_from_xdr_buf(struct xdr_buf *, unsigned int, void *, unsigned int); extern int write_bytes_to_xdr_buf(struct xdr_buf *, unsigned int, void *, unsigned int); -- cgit v1.2.3 From d7242c4641fba521a1ea9dbccb11a40cf38cd912 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 18 Mar 2020 17:22:47 -0400 Subject: pNFS: Add a helper to allocate the array of buckets Signed-off-by: Trond Myklebust --- include/linux/nfs_xdr.h | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) (limited to 'include') diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 94c77ed55ce1..e91c917c9c1c 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1270,10 +1270,19 @@ struct pnfs_commit_bucket { struct nfs_writeverf direct_verf; }; +struct pnfs_commit_array { + struct list_head cinfo_list; + struct list_head lseg_list; + struct pnfs_layout_segment *lseg; + struct rcu_head rcu; + unsigned int nbuckets; + struct pnfs_commit_bucket buckets[]; +}; + struct pnfs_ds_commit_info { - int nwritten; - int ncommitting; - int nbuckets; + unsigned int nwritten; + unsigned int ncommitting; + unsigned int nbuckets; struct pnfs_commit_bucket *buckets; }; -- cgit v1.2.3 From 62a89501a3bde310ba339cc90bb05879528d84a0 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Wed, 12 Feb 2020 11:12:35 -0500 Subject: xprtrdma: Enhance MR-related trace points Two changes: - Show the number of SG entries that were mapped. This helps debug DMA-related problems. - Record the MR's resource ID instead of its memory address. This groups each MR with its associated rdma-tool output, and reduces needless exposure of memory addresses. Signed-off-by: Chuck Lever Signed-off-by: Anna Schumaker --- include/trace/events/rpcrdma.h | 56 ++++++++++++++++++++++-------------------- 1 file changed, 30 insertions(+), 26 deletions(-) (limited to 'include') diff --git a/include/trace/events/rpcrdma.h b/include/trace/events/rpcrdma.h index c0e4c93324f5..87f4461ab108 100644 --- a/include/trace/events/rpcrdma.h +++ b/include/trace/events/rpcrdma.h @@ -228,20 +228,20 @@ DECLARE_EVENT_CLASS(xprtrdma_frwr_done, TP_ARGS(wc, frwr), TP_STRUCT__entry( - __field(const void *, mr) + __field(u32, mr_id) __field(unsigned int, status) __field(unsigned int, vendor_err) ), TP_fast_assign( - __entry->mr = container_of(frwr, struct rpcrdma_mr, frwr); + __entry->mr_id = frwr->fr_mr->res.id; __entry->status = wc->status; __entry->vendor_err = __entry->status ? wc->vendor_err : 0; ), TP_printk( - "mr=%p: %s (%u/0x%x)", - __entry->mr, rdma_show_wc_status(__entry->status), + "mr.id=%u: %s (%u/0x%x)", + __entry->mr_id, rdma_show_wc_status(__entry->status), __entry->status, __entry->vendor_err ) ); @@ -274,7 +274,8 @@ DECLARE_EVENT_CLASS(xprtrdma_mr, TP_ARGS(mr), TP_STRUCT__entry( - __field(const void *, mr) + __field(u32, mr_id) + __field(int, nents) __field(u32, handle) __field(u32, length) __field(u64, offset) @@ -282,15 +283,16 @@ DECLARE_EVENT_CLASS(xprtrdma_mr, ), TP_fast_assign( - __entry->mr = mr; + __entry->mr_id = mr->frwr.fr_mr->res.id; + __entry->nents = mr->mr_nents; __entry->handle = mr->mr_handle; __entry->length = mr->mr_length; __entry->offset = mr->mr_offset; __entry->dir = mr->mr_dir; ), - TP_printk("mr=%p %u@0x%016llx:0x%08x (%s)", - __entry->mr, __entry->length, + TP_printk("mr.id=%u nents=%d %u@0x%016llx:0x%08x (%s)", + __entry->mr_id, __entry->nents, __entry->length, (unsigned long long)__entry->offset, __entry->handle, xprtrdma_show_direction(__entry->dir) ) @@ -920,17 +922,17 @@ TRACE_EVENT(xprtrdma_frwr_alloc, TP_ARGS(mr, rc), TP_STRUCT__entry( - __field(const void *, mr) + __field(u32, mr_id) __field(int, rc) ), TP_fast_assign( - __entry->mr = mr; - __entry->rc = rc; + __entry->mr_id = mr->frwr.fr_mr->res.id; + __entry->rc = rc; ), - TP_printk("mr=%p: rc=%d", - __entry->mr, __entry->rc + TP_printk("mr.id=%u: rc=%d", + __entry->mr_id, __entry->rc ) ); @@ -943,7 +945,8 @@ TRACE_EVENT(xprtrdma_frwr_dereg, TP_ARGS(mr, rc), TP_STRUCT__entry( - __field(const void *, mr) + __field(u32, mr_id) + __field(int, nents) __field(u32, handle) __field(u32, length) __field(u64, offset) @@ -952,7 +955,8 @@ TRACE_EVENT(xprtrdma_frwr_dereg, ), TP_fast_assign( - __entry->mr = mr; + __entry->mr_id = mr->frwr.fr_mr->res.id; + __entry->nents = mr->mr_nents; __entry->handle = mr->mr_handle; __entry->length = mr->mr_length; __entry->offset = mr->mr_offset; @@ -960,8 +964,8 @@ TRACE_EVENT(xprtrdma_frwr_dereg, __entry->rc = rc; ), - TP_printk("mr=%p %u@0x%016llx:0x%08x (%s): rc=%d", - __entry->mr, __entry->length, + TP_printk("mr.id=%u nents=%d %u@0x%016llx:0x%08x (%s): rc=%d", + __entry->mr_id, __entry->nents, __entry->length, (unsigned long long)__entry->offset, __entry->handle, xprtrdma_show_direction(__entry->dir), __entry->rc @@ -977,21 +981,21 @@ TRACE_EVENT(xprtrdma_frwr_sgerr, TP_ARGS(mr, sg_nents), TP_STRUCT__entry( - __field(const void *, mr) + __field(u32, mr_id) __field(u64, addr) __field(u32, dir) __field(int, nents) ), TP_fast_assign( - __entry->mr = mr; + __entry->mr_id = mr->frwr.fr_mr->res.id; __entry->addr = mr->mr_sg->dma_address; __entry->dir = mr->mr_dir; __entry->nents = sg_nents; ), - TP_printk("mr=%p dma addr=0x%llx (%s) sg_nents=%d", - __entry->mr, __entry->addr, + TP_printk("mr.id=%u DMA addr=0x%llx (%s) sg_nents=%d", + __entry->mr_id, __entry->addr, xprtrdma_show_direction(__entry->dir), __entry->nents ) @@ -1006,7 +1010,7 @@ TRACE_EVENT(xprtrdma_frwr_maperr, TP_ARGS(mr, num_mapped), TP_STRUCT__entry( - __field(const void *, mr) + __field(u32, mr_id) __field(u64, addr) __field(u32, dir) __field(int, num_mapped) @@ -1014,15 +1018,15 @@ TRACE_EVENT(xprtrdma_frwr_maperr, ), TP_fast_assign( - __entry->mr = mr; + __entry->mr_id = mr->frwr.fr_mr->res.id; __entry->addr = mr->mr_sg->dma_address; __entry->dir = mr->mr_dir; __entry->num_mapped = num_mapped; __entry->nents = mr->mr_nents; ), - TP_printk("mr=%p dma addr=0x%llx (%s) nents=%d of %d", - __entry->mr, __entry->addr, + TP_printk("mr.id=%u DMA addr=0x%llx (%s) nents=%d of %d", + __entry->mr_id, __entry->addr, xprtrdma_show_direction(__entry->dir), __entry->num_mapped, __entry->nents ) @@ -1031,7 +1035,7 @@ TRACE_EVENT(xprtrdma_frwr_maperr, DEFINE_MR_EVENT(localinv); DEFINE_MR_EVENT(map); DEFINE_MR_EVENT(unmap); -DEFINE_MR_EVENT(remoteinv); +DEFINE_MR_EVENT(reminv); DEFINE_MR_EVENT(recycle); TRACE_EVENT(xprtrdma_dma_maperr, -- cgit v1.2.3 From 81fe0c57f4e136375f3bcda74af413f82a34a1bb Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 21 Feb 2020 17:00:38 -0500 Subject: xprtrdma: Invoke rpcrdma_ia_open in the connect worker Move rdma_cm_id creation into rpcrdma_ep_create() so that it is now responsible for allocating all per-connection hardware resources. With this clean-up, all three arms of the switch statement in rpcrdma_ep_connect are exactly the same now, thus the switch can be removed. Because device removal behaves a little differently than disconnection, there is a little more work to be done before rpcrdma_ep_destroy() can release the connection's rdma_cm_id. So it is not quite symmetrical with rpcrdma_ep_create() yet. Signed-off-by: Chuck Lever Signed-off-by: Anna Schumaker --- include/trace/events/rpcrdma.h | 1 - 1 file changed, 1 deletion(-) (limited to 'include') diff --git a/include/trace/events/rpcrdma.h b/include/trace/events/rpcrdma.h index 87f4461ab108..ba37c47b51e8 100644 --- a/include/trace/events/rpcrdma.h +++ b/include/trace/events/rpcrdma.h @@ -415,7 +415,6 @@ DEFINE_CONN_EVENT(disconnect); DEFINE_RXPRT_EVENT(xprtrdma_create); DEFINE_RXPRT_EVENT(xprtrdma_op_destroy); DEFINE_RXPRT_EVENT(xprtrdma_remove); -DEFINE_RXPRT_EVENT(xprtrdma_reinsert); DEFINE_RXPRT_EVENT(xprtrdma_op_inject_dsc); DEFINE_RXPRT_EVENT(xprtrdma_op_close); DEFINE_RXPRT_EVENT(xprtrdma_op_setport); -- cgit v1.2.3 From d6ccebf956338ea015d7d54c4a4c9c17605707cb Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 21 Feb 2020 17:00:49 -0500 Subject: xprtrdma: Disconnect on flushed completion Completion errors after a disconnect often occur much sooner than a CM_DISCONNECT event. Use this to try to detect connection loss more quickly. Note that other kernel ULPs do take care to disconnect explicitly when a WR is flushed. Signed-off-by: Chuck Lever Signed-off-by: Anna Schumaker --- include/trace/events/rpcrdma.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/trace/events/rpcrdma.h b/include/trace/events/rpcrdma.h index ba37c47b51e8..cfbe28ad2614 100644 --- a/include/trace/events/rpcrdma.h +++ b/include/trace/events/rpcrdma.h @@ -109,7 +109,7 @@ DECLARE_EVENT_CLASS(xprtrdma_connect_class, __assign_str(port, rpcrdma_portstr(r_xprt)); ), - TP_printk("peer=[%s]:%s r_xprt=%p: rc=%d connect status=%d", + TP_printk("peer=[%s]:%s r_xprt=%p: rc=%d connection status=%d", __get_str(addr), __get_str(port), __entry->r_xprt, __entry->rc, __entry->connect_status ) @@ -411,6 +411,7 @@ TRACE_EVENT(xprtrdma_inline_thresh, DEFINE_CONN_EVENT(connect); DEFINE_CONN_EVENT(disconnect); +DEFINE_CONN_EVENT(flush_dct); DEFINE_RXPRT_EVENT(xprtrdma_create); DEFINE_RXPRT_EVENT(xprtrdma_op_destroy); -- cgit v1.2.3 From 93aa8e0a9de80e1df2be17158a3469285e572b39 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 21 Feb 2020 17:00:54 -0500 Subject: xprtrdma: Merge struct rpcrdma_ia into struct rpcrdma_ep I eventually want to allocate rpcrdma_ep separately from struct rpcrdma_xprt so that on occasion there can be more than one ep per xprt. The new struct rpcrdma_ep will contain all the fields currently in rpcrdma_ia and in rpcrdma_ep. This is all the device and CM settings for the connection, in addition to per-connection settings negotiated with the remote. Take this opportunity to rename the existing ep fields from rep_* to re_* to disambiguate these from struct rpcrdma_rep. Signed-off-by: Chuck Lever Signed-off-by: Anna Schumaker --- include/trace/events/rpcrdma.h | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'include') diff --git a/include/trace/events/rpcrdma.h b/include/trace/events/rpcrdma.h index cfbe28ad2614..843269f0e291 100644 --- a/include/trace/events/rpcrdma.h +++ b/include/trace/events/rpcrdma.h @@ -104,7 +104,7 @@ DECLARE_EVENT_CLASS(xprtrdma_connect_class, TP_fast_assign( __entry->r_xprt = r_xprt; __entry->rc = rc; - __entry->connect_status = r_xprt->rx_ep.rep_connected; + __entry->connect_status = r_xprt->rx_ep.re_connect_status; __assign_str(addr, rpcrdma_addrstr(r_xprt)); __assign_str(port, rpcrdma_portstr(r_xprt)); ), @@ -394,10 +394,10 @@ TRACE_EVENT(xprtrdma_inline_thresh, const struct rpcrdma_ep *ep = &r_xprt->rx_ep; __entry->r_xprt = r_xprt; - __entry->inline_send = ep->rep_inline_send; - __entry->inline_recv = ep->rep_inline_recv; - __entry->max_send = ep->rep_max_inline_send; - __entry->max_recv = ep->rep_max_inline_recv; + __entry->inline_send = ep->re_inline_send; + __entry->inline_recv = ep->re_inline_recv; + __entry->max_send = ep->re_max_inline_send; + __entry->max_recv = ep->re_max_inline_recv; __assign_str(addr, rpcrdma_addrstr(r_xprt)); __assign_str(port, rpcrdma_portstr(r_xprt)); ), @@ -803,7 +803,7 @@ TRACE_EVENT(xprtrdma_post_recvs, __entry->r_xprt = r_xprt; __entry->count = count; __entry->status = status; - __entry->posted = r_xprt->rx_ep.rep_receive_count; + __entry->posted = r_xprt->rx_ep.re_receive_count; __assign_str(addr, rpcrdma_addrstr(r_xprt)); __assign_str(port, rpcrdma_portstr(r_xprt)); ), -- cgit v1.2.3 From 745b734c9bb80559b8ca64ae688901afefc1c3fd Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 21 Feb 2020 17:01:00 -0500 Subject: xprtrdma: Extract sockaddr from struct rdma_cm_id rpcrdma_cm_event_handler() is always passed an @id pointer that is valid. However, in a subsequent patch, we won't be able to extract an r_xprt in every case. So instead of using the r_xprt's presentation address strings, extract them from struct rdma_cm_id. Signed-off-by: Chuck Lever Signed-off-by: Anna Schumaker --- include/trace/events/rpcrdma.h | 78 ++++++++++++++++++++++++++++-------------- 1 file changed, 53 insertions(+), 25 deletions(-) (limited to 'include') diff --git a/include/trace/events/rpcrdma.h b/include/trace/events/rpcrdma.h index 843269f0e291..295f75b9b796 100644 --- a/include/trace/events/rpcrdma.h +++ b/include/trace/events/rpcrdma.h @@ -375,47 +375,74 @@ TRACE_EVENT(xprtrdma_cm_event, TRACE_EVENT(xprtrdma_inline_thresh, TP_PROTO( - const struct rpcrdma_xprt *r_xprt + const struct rpcrdma_ep *ep ), - TP_ARGS(r_xprt), + TP_ARGS(ep), TP_STRUCT__entry( - __field(const void *, r_xprt) __field(unsigned int, inline_send) __field(unsigned int, inline_recv) __field(unsigned int, max_send) __field(unsigned int, max_recv) - __string(addr, rpcrdma_addrstr(r_xprt)) - __string(port, rpcrdma_portstr(r_xprt)) + __array(unsigned char, srcaddr, sizeof(struct sockaddr_in6)) + __array(unsigned char, dstaddr, sizeof(struct sockaddr_in6)) ), TP_fast_assign( - const struct rpcrdma_ep *ep = &r_xprt->rx_ep; + const struct rdma_cm_id *id = ep->re_id; - __entry->r_xprt = r_xprt; __entry->inline_send = ep->re_inline_send; __entry->inline_recv = ep->re_inline_recv; __entry->max_send = ep->re_max_inline_send; __entry->max_recv = ep->re_max_inline_recv; - __assign_str(addr, rpcrdma_addrstr(r_xprt)); - __assign_str(port, rpcrdma_portstr(r_xprt)); + memcpy(__entry->srcaddr, &id->route.addr.src_addr, + sizeof(struct sockaddr_in6)); + memcpy(__entry->dstaddr, &id->route.addr.dst_addr, + sizeof(struct sockaddr_in6)); ), - TP_printk("peer=[%s]:%s r_xprt=%p neg send/recv=%u/%u, calc send/recv=%u/%u", - __get_str(addr), __get_str(port), __entry->r_xprt, + TP_printk("%pISpc -> %pISpc neg send/recv=%u/%u, calc send/recv=%u/%u", + __entry->srcaddr, __entry->dstaddr, __entry->inline_send, __entry->inline_recv, __entry->max_send, __entry->max_recv ) ); +TRACE_EVENT(xprtrdma_remove, + TP_PROTO( + const struct rpcrdma_ep *ep + ), + + TP_ARGS(ep), + + TP_STRUCT__entry( + __array(unsigned char, srcaddr, sizeof(struct sockaddr_in6)) + __array(unsigned char, dstaddr, sizeof(struct sockaddr_in6)) + __string(name, ep->re_id->device->name) + ), + + TP_fast_assign( + const struct rdma_cm_id *id = ep->re_id; + + memcpy(__entry->srcaddr, &id->route.addr.src_addr, + sizeof(struct sockaddr_in6)); + memcpy(__entry->dstaddr, &id->route.addr.dst_addr, + sizeof(struct sockaddr_in6)); + __assign_str(name, id->device->name); + ), + + TP_printk("%pISpc -> %pISpc device=%s", + __entry->srcaddr, __entry->dstaddr, __get_str(name) + ) +); + DEFINE_CONN_EVENT(connect); DEFINE_CONN_EVENT(disconnect); DEFINE_CONN_EVENT(flush_dct); DEFINE_RXPRT_EVENT(xprtrdma_create); DEFINE_RXPRT_EVENT(xprtrdma_op_destroy); -DEFINE_RXPRT_EVENT(xprtrdma_remove); DEFINE_RXPRT_EVENT(xprtrdma_op_inject_dsc); DEFINE_RXPRT_EVENT(xprtrdma_op_close); DEFINE_RXPRT_EVENT(xprtrdma_op_setport); @@ -482,32 +509,33 @@ TRACE_EVENT(xprtrdma_op_set_cto, TRACE_EVENT(xprtrdma_qp_event, TP_PROTO( - const struct rpcrdma_xprt *r_xprt, + const struct rpcrdma_ep *ep, const struct ib_event *event ), - TP_ARGS(r_xprt, event), + TP_ARGS(ep, event), TP_STRUCT__entry( - __field(const void *, r_xprt) - __field(unsigned int, event) + __field(unsigned long, event) __string(name, event->device->name) - __string(addr, rpcrdma_addrstr(r_xprt)) - __string(port, rpcrdma_portstr(r_xprt)) + __array(unsigned char, srcaddr, sizeof(struct sockaddr_in6)) + __array(unsigned char, dstaddr, sizeof(struct sockaddr_in6)) ), TP_fast_assign( - __entry->r_xprt = r_xprt; + const struct rdma_cm_id *id = ep->re_id; + __entry->event = event->event; __assign_str(name, event->device->name); - __assign_str(addr, rpcrdma_addrstr(r_xprt)); - __assign_str(port, rpcrdma_portstr(r_xprt)); + memcpy(__entry->srcaddr, &id->route.addr.src_addr, + sizeof(struct sockaddr_in6)); + memcpy(__entry->dstaddr, &id->route.addr.dst_addr, + sizeof(struct sockaddr_in6)); ), - TP_printk("peer=[%s]:%s r_xprt=%p: dev %s: %s (%u)", - __get_str(addr), __get_str(port), __entry->r_xprt, - __get_str(name), rdma_show_ib_event(__entry->event), - __entry->event + TP_printk("%pISpc -> %pISpc device=%s %s (%lu)", + __entry->srcaddr, __entry->dstaddr, __get_str(name), + rdma_show_ib_event(__entry->event), __entry->event ) ); -- cgit v1.2.3 From e28ce90083f032ca0e8ea03478f5b6a38f5930f7 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 21 Feb 2020 17:01:05 -0500 Subject: xprtrdma: kmalloc rpcrdma_ep separate from rpcrdma_xprt Change the rpcrdma_xprt_disconnect() function so that it no longer waits for the DISCONNECTED event. This prevents blocking if the remote is unresponsive. In rpcrdma_xprt_disconnect(), the transport's rpcrdma_ep is detached. Upon return from rpcrdma_xprt_disconnect(), the transport (r_xprt) is ready immediately for a new connection. The RDMA_CM_DEVICE_REMOVAL and RDMA_CM_DISCONNECTED events are now handled almost identically. However, because the lifetimes of rpcrdma_xprt structures and rpcrdma_ep structures are now independent, creating an rpcrdma_ep needs to take a module ref count. The ep now owns most of the hardware resources for a transport. Also, a kref is needed to ensure that rpcrdma_ep sticks around long enough for the cm_event_handler to finish. Signed-off-by: Chuck Lever Signed-off-by: Anna Schumaker --- include/trace/events/rpcrdma.h | 63 ++---------------------------------------- 1 file changed, 2 insertions(+), 61 deletions(-) (limited to 'include') diff --git a/include/trace/events/rpcrdma.h b/include/trace/events/rpcrdma.h index 295f75b9b796..81b87428f166 100644 --- a/include/trace/events/rpcrdma.h +++ b/include/trace/events/rpcrdma.h @@ -104,7 +104,7 @@ DECLARE_EVENT_CLASS(xprtrdma_connect_class, TP_fast_assign( __entry->r_xprt = r_xprt; __entry->rc = rc; - __entry->connect_status = r_xprt->rx_ep.re_connect_status; + __entry->connect_status = r_xprt->rx_ep->re_connect_status; __assign_str(addr, rpcrdma_addrstr(r_xprt)); __assign_str(port, rpcrdma_portstr(r_xprt)); ), @@ -342,37 +342,6 @@ DECLARE_EVENT_CLASS(xprtrdma_cb_event, ** Connection events **/ -TRACE_EVENT(xprtrdma_cm_event, - TP_PROTO( - const struct rpcrdma_xprt *r_xprt, - struct rdma_cm_event *event - ), - - TP_ARGS(r_xprt, event), - - TP_STRUCT__entry( - __field(const void *, r_xprt) - __field(unsigned int, event) - __field(int, status) - __string(addr, rpcrdma_addrstr(r_xprt)) - __string(port, rpcrdma_portstr(r_xprt)) - ), - - TP_fast_assign( - __entry->r_xprt = r_xprt; - __entry->event = event->event; - __entry->status = event->status; - __assign_str(addr, rpcrdma_addrstr(r_xprt)); - __assign_str(port, rpcrdma_portstr(r_xprt)); - ), - - TP_printk("peer=[%s]:%s r_xprt=%p: %s (%u/%d)", - __get_str(addr), __get_str(port), - __entry->r_xprt, rdma_show_cm_event(__entry->event), - __entry->event, __entry->status - ) -); - TRACE_EVENT(xprtrdma_inline_thresh, TP_PROTO( const struct rpcrdma_ep *ep @@ -409,34 +378,6 @@ TRACE_EVENT(xprtrdma_inline_thresh, ) ); -TRACE_EVENT(xprtrdma_remove, - TP_PROTO( - const struct rpcrdma_ep *ep - ), - - TP_ARGS(ep), - - TP_STRUCT__entry( - __array(unsigned char, srcaddr, sizeof(struct sockaddr_in6)) - __array(unsigned char, dstaddr, sizeof(struct sockaddr_in6)) - __string(name, ep->re_id->device->name) - ), - - TP_fast_assign( - const struct rdma_cm_id *id = ep->re_id; - - memcpy(__entry->srcaddr, &id->route.addr.src_addr, - sizeof(struct sockaddr_in6)); - memcpy(__entry->dstaddr, &id->route.addr.dst_addr, - sizeof(struct sockaddr_in6)); - __assign_str(name, id->device->name); - ), - - TP_printk("%pISpc -> %pISpc device=%s", - __entry->srcaddr, __entry->dstaddr, __get_str(name) - ) -); - DEFINE_CONN_EVENT(connect); DEFINE_CONN_EVENT(disconnect); DEFINE_CONN_EVENT(flush_dct); @@ -831,7 +772,7 @@ TRACE_EVENT(xprtrdma_post_recvs, __entry->r_xprt = r_xprt; __entry->count = count; __entry->status = status; - __entry->posted = r_xprt->rx_ep.re_receive_count; + __entry->posted = r_xprt->rx_ep->re_receive_count; __assign_str(addr, rpcrdma_addrstr(r_xprt)); __assign_str(port, rpcrdma_portstr(r_xprt)); ), -- cgit v1.2.3 From c21e7168848d4ff4158120dbd4464f0d5cfb1456 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 19 Mar 2020 13:36:36 -0400 Subject: NFSv4/pnfs: Support a list of commit arrays in struct pnfs_ds_commit_info When we have multiple layout segments with different lists of mirrored data, we need to track the commits on a per layout segment basis. This patch adds a list to support this tracking in struct pnfs_ds_commit_info. Signed-off-by: Trond Myklebust --- include/linux/nfs_xdr.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index e91c917c9c1c..9946787eda72 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1280,6 +1280,7 @@ struct pnfs_commit_array { }; struct pnfs_ds_commit_info { + struct list_head commits; unsigned int nwritten; unsigned int ncommitting; unsigned int nbuckets; -- cgit v1.2.3 From a9901899b649dc80ef75c14d6d78059cae14def7 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 20 Mar 2020 16:04:06 -0400 Subject: pNFS: Add infrastructure for cleaning up per-layout commit structures Ensure that both the file and flexfiles layout types clean up when freeing the layout segments. Signed-off-by: Trond Myklebust --- include/linux/nfs_xdr.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 9946787eda72..33be2ee2a248 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1275,6 +1275,7 @@ struct pnfs_commit_array { struct list_head lseg_list; struct pnfs_layout_segment *lseg; struct rcu_head rcu; + refcount_t refcount; unsigned int nbuckets; struct pnfs_commit_bucket buckets[]; }; -- cgit v1.2.3 From 0aa647b7369dd29de0789c321111b2e4668c46b2 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 21 Mar 2020 09:50:05 -0400 Subject: NFS: Remove bucket array from struct pnfs_ds_commit_info Remove the unused bucket array in struct pnfs_ds_commit_info. Signed-off-by: Trond Myklebust --- include/linux/nfs_xdr.h | 13 ------------- 1 file changed, 13 deletions(-) (limited to 'include') diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 33be2ee2a248..2903597ec88c 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1284,8 +1284,6 @@ struct pnfs_ds_commit_info { struct list_head commits; unsigned int nwritten; unsigned int ncommitting; - unsigned int nbuckets; - struct pnfs_commit_bucket *buckets; }; struct nfs41_state_protection { @@ -1396,22 +1394,11 @@ struct nfs41_free_stateid_res { unsigned int status; }; -static inline void -nfs_free_pnfs_ds_cinfo(struct pnfs_ds_commit_info *cinfo) -{ - kfree(cinfo->buckets); -} - #else struct pnfs_ds_commit_info { }; -static inline void -nfs_free_pnfs_ds_cinfo(struct pnfs_ds_commit_info *cinfo) -{ -} - #endif /* CONFIG_NFS_V4_1 */ #ifdef CONFIG_NFS_V4_2 -- cgit v1.2.3 From 9c455a8c1e146dac3a6d1405fe6a7096177b9546 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 21 Mar 2020 11:13:05 -0400 Subject: NFS/pNFS: Clean up pNFS commit operations Move the pNFS commit related operations into a separate structure that can be carried by the pnfs_ds_commit_info. Signed-off-by: Trond Myklebust --- include/linux/nfs_xdr.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 2903597ec88c..adbbeae9ce5b 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1284,6 +1284,7 @@ struct pnfs_ds_commit_info { struct list_head commits; unsigned int nwritten; unsigned int ncommitting; + const struct pnfs_commit_ops *ops; }; struct nfs41_state_protection { -- cgit v1.2.3 From c84bea59449aaa699a0600a50f59d441cc1d4501 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 22 Mar 2020 14:47:38 -0400 Subject: NFS/pNFS: Simplify bucket layout segment reference counting Signed-off-by: Trond Myklebust --- include/linux/nfs_xdr.h | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index adbbeae9ce5b..7bbb1f6fc1b1 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1265,8 +1265,7 @@ struct nfstime4 { struct pnfs_commit_bucket { struct list_head written; struct list_head committing; - struct pnfs_layout_segment *wlseg; - struct pnfs_layout_segment *clseg; + struct pnfs_layout_segment *lseg; struct nfs_writeverf direct_verf; }; -- cgit v1.2.3 From 08ca8b21f760c0ed5034a5c122092eec22ccf8f4 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 1 Apr 2020 13:04:49 -0400 Subject: NFS: Fix races nfs_page_group_destroy() vs nfs_destroy_unlinked_subrequests() When a subrequest is being detached from the subgroup, we want to ensure that it is not holding the group lock, or in the process of waiting for the group lock. Fixes: 5b2b5187fa85 ("NFS: Fix nfs_page_group_destroy() and nfs_lock_and_join_requests() race cases") Signed-off-by: Trond Myklebust --- include/linux/nfs_page.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/linux/nfs_page.h b/include/linux/nfs_page.h index 0bbd587fac6a..7e9419d74b86 100644 --- a/include/linux/nfs_page.h +++ b/include/linux/nfs_page.h @@ -142,6 +142,8 @@ extern void nfs_unlock_and_release_request(struct nfs_page *); extern int nfs_page_group_lock(struct nfs_page *); extern void nfs_page_group_unlock(struct nfs_page *); extern bool nfs_page_group_sync_on_bit(struct nfs_page *, unsigned int); +extern int nfs_page_set_headlock(struct nfs_page *req); +extern void nfs_page_clear_headlock(struct nfs_page *req); extern bool nfs_async_iocounter_wait(struct rpc_task *, struct nfs_lock_context *); /* -- cgit v1.2.3 From a62f8e3bd836bf1abde1648a45e14afd050dbd23 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 30 Mar 2020 11:12:16 -0400 Subject: NFS: Clean up nfs_lock_and_join_requests() Clean up nfs_lock_and_join_requests() to simplify the calculation of the range covered by the page group, taking into account the presence of mirrors. Signed-off-by: Trond Myklebust --- include/linux/nfs_page.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/nfs_page.h b/include/linux/nfs_page.h index 7e9419d74b86..dd205bc6bc58 100644 --- a/include/linux/nfs_page.h +++ b/include/linux/nfs_page.h @@ -139,6 +139,7 @@ extern size_t nfs_generic_pg_test(struct nfs_pageio_descriptor *desc, extern int nfs_wait_on_request(struct nfs_page *); extern void nfs_unlock_request(struct nfs_page *req); extern void nfs_unlock_and_release_request(struct nfs_page *); +extern int nfs_page_group_lock_subrequests(struct nfs_page *head); extern int nfs_page_group_lock(struct nfs_page *); extern void nfs_page_group_unlock(struct nfs_page *); extern bool nfs_page_group_sync_on_bit(struct nfs_page *, unsigned int); -- cgit v1.2.3 From e00ed89d7bd59c4ae49d6aeeee567187b1357a4b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 30 Mar 2020 12:40:47 -0400 Subject: NFS: Refactor nfs_lock_and_join_requests() Refactor nfs_lock_and_join_requests() in order to separate out the subrequest merging into its own function nfs_lock_and_join_group() that can be used by O_DIRECT. Signed-off-by: Trond Myklebust --- include/linux/nfs_page.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/nfs_page.h b/include/linux/nfs_page.h index dd205bc6bc58..99198c039bd6 100644 --- a/include/linux/nfs_page.h +++ b/include/linux/nfs_page.h @@ -139,6 +139,7 @@ extern size_t nfs_generic_pg_test(struct nfs_pageio_descriptor *desc, extern int nfs_wait_on_request(struct nfs_page *); extern void nfs_unlock_request(struct nfs_page *req); extern void nfs_unlock_and_release_request(struct nfs_page *); +extern struct nfs_page *nfs_page_group_lock_head(struct nfs_page *req); extern int nfs_page_group_lock_subrequests(struct nfs_page *head); extern int nfs_page_group_lock(struct nfs_page *); extern void nfs_page_group_unlock(struct nfs_page *); -- cgit v1.2.3 From ed5d588fe47feef290f271022820e255d8371561 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 30 Mar 2020 20:57:49 -0400 Subject: NFS: Try to join page groups before an O_DIRECT retransmission If we have to retransmit requests, try to join their page groups first. Signed-off-by: Trond Myklebust --- include/linux/nfs_page.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/nfs_page.h b/include/linux/nfs_page.h index 99198c039bd6..c32c15216da3 100644 --- a/include/linux/nfs_page.h +++ b/include/linux/nfs_page.h @@ -141,6 +141,7 @@ extern void nfs_unlock_request(struct nfs_page *req); extern void nfs_unlock_and_release_request(struct nfs_page *); extern struct nfs_page *nfs_page_group_lock_head(struct nfs_page *req); extern int nfs_page_group_lock_subrequests(struct nfs_page *head); +extern void nfs_join_page_group(struct nfs_page *head, struct inode *inode); extern int nfs_page_group_lock(struct nfs_page *); extern void nfs_page_group_unlock(struct nfs_page *); extern bool nfs_page_group_sync_on_bit(struct nfs_page *, unsigned int); -- cgit v1.2.3 From 93ce4af774bc3d8a72ce2271d03241c96383629d Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 6 Apr 2020 13:39:29 -0400 Subject: NFS: Clean up process of marking inode stale. Instead of the various open coded calls to set the NFS_INO_STALE bit and call nfs_zap_caches(), consolidate them into a single function nfs_set_inode_stale(). Signed-off-by: Trond Myklebust --- include/linux/nfs_fs.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 5d5b91e54f73..73eda45f1cfd 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -354,6 +354,7 @@ static inline unsigned long nfs_save_change_attribute(struct inode *dir) extern int nfs_sync_mapping(struct address_space *mapping); extern void nfs_zap_mapping(struct inode *inode, struct address_space *mapping); extern void nfs_zap_caches(struct inode *); +extern void nfs_set_inode_stale(struct inode *inode); extern void nfs_invalidate_atime(struct inode *); extern struct inode *nfs_fhget(struct super_block *, struct nfs_fh *, struct nfs_fattr *, struct nfs4_label *); -- cgit v1.2.3